Import Packages

import numpy as np
import cv2
import time
import matplotlib.pyplot as plt

%matplotlib inline

Import python File

%load_ext autoreload
#using autoreload 1 to reload all modules imported by %aimport everytime before executing the python code typed .
%autoreload 2

%aimport lane_detection, car_detection, main

Lane Detection Program

Testing sobel filter

img = cv2.imread('./test_images/test5.jpg')

img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)

sobel_img = lane_detection.abs_sobel_thresh(gray, thresh=(10, 200))

lane_detection.plt_images(img, 'Source image', sobel_img, 'Sobel image')

dir_binary = lane_detection.dir_threshold(gray, thresh=(np.pi/6, np.pi/2))

lane_detection.plt_images(img, 'Source image', dir_binary, 'dir Sobel image')

combined = ((sobel_img == 1) & (dir_binary == 1))
lane_detection.plt_images(sobel_img, 'sobel image', combined, 'combined Sobel image')

thresholded = lane_detection.combine_thresh(img,s_thresh=(100, 255), l_thresh=(120, 255))

lane_detection.plt_images(img, 'Source image', thresholded, 'combined threshold image')

Testing perspective warp

thresholded_img, warped_img, Minv = lane_detection.warp_image_to_birdseye_view_gray(img)

lane_detection.plt_images(img, 'Source image', warped_img, 'warped image')

Testing sliding window

left_fit, right_fit,out_img, lefty, leftx, righty, rightx, ploty = lane_detection.fitlines(warped_img, nwindows=15, margin=100, minpix = 50 )
plt.imshow(out_img)

<matplotlib.image.AxesImage at 0x1f6c2b02b80>

Testing vehicle position and radius of curvature

center,left_curverad, right_curverad = lane_detection.lane_curvatures(img, lefty, leftx, righty, rightx, ploty)
print(center)
print(left_curverad)
print(right_curverad)

0.29517596785131356
1570.2880410982805
5567.999760624177

Testing of draw lanes

result,color_warp = lane_detection.draw_lane(img, warped_img, left_fit, right_fit, ploty, center, left_curverad, right_curverad, Minv)
lane_detection.plt_images(result, 'Output image', color_warp, 'Lane image')

Testing laneDetection_pipeline_v1 function

processed_img = lane_detection.laneDetection_pipeline_v1(img)
plt.figure(figsize=(10,5))
plt.imshow(processed_img)

<matplotlib.image.AxesImage at 0x1f6c09502e0>

Testing laneDetection_pipeline_v2 function

processed_img = lane_detection.laneDetection_pipeline_v2(img)
plt.figure(figsize=(10,5))
plt.imshow(processed_img)

<matplotlib.image.AxesImage at 0x1f6bf5eb9d0>

Car Detection Program

Load yolo weights and layers

net = car_detection.getNet()

layers_name = car_detection.loadWeightsLayers(net)

print("yolo weights and layers are loaded.")
print("Layer names: ", layers_name)

yolo weights and layers are loaded.
Layer names:  ['yolo_82', 'yolo_94', 'yolo_106']

Run the interface on the test image

start_t = time.time()

layers_output = car_detection.netForwardOutput(img, net, layers_name)

print("Interface took: {}".format(time.time() - start_t))

Interface took: 1.0669240951538086

Draw boxes on the detected car

boxes = []
confidences = []
classIDs = []
idxs=[]

idxs, boxes, confidences = car_detection.getBoxes(layers_output, img)
img = car_detection.drawBoxes(idxs, boxes, confidences, img)

plt.imshow(img)

<matplotlib.image.AxesImage at 0x1f6bef40a30>

Draw detected cars on the left

img = cv2.imread('./test_images/test5.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

boxes = []
confidences = []
classIDs = []
idxs=[]

idxs, boxes, confidences = car_detection.getBoxes(layers_output, img)
img = car_detection.draw_cars(idxs,boxes,img)
img = car_detection.drawBoxes(idxs, boxes, confidences, img)

plt.imshow(img)

<matplotlib.image.AxesImage at 0x1f6bf6b8250>

Testing whole Program

Normal mode

img = cv2.imread('./test_images/test5.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
processed_img_whole = main.pipeline_v1(img)
plt.imshow(processed_img_whole)

<matplotlib.image.AxesImage at 0x1f6c0d78310>

Normal mode with draw cars

img = cv2.imread('./test_images/test5.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
processed_img_whole = main.pipeline_v2(img)
plt.imshow(processed_img_whole)

<matplotlib.image.AxesImage at 0x1f6bf3007f0>

Debug mode

img = cv2.imread('./test_images/test5.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
processed_img_whole = main.pipeline_v3(img)
plt.imshow(processed_img_whole)

<matplotlib.image.AxesImage at 0x1f6bf2cfc40>

Debug mode with draw cars

img = cv2.imread('./test_images/test5.jpg')
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
processed_img_whole = main.pipeline_v4(img)
plt.imshow(processed_img_whole)

<matplotlib.image.AxesImage at 0x1f6c08debb0>

Creating the output video

main.createVideo("./test_videos/challenge_video.mp4","challenge_video_output.mp4")

main.createVideo("./test_videos/challenge_video.mp4","challenge_video_output_debug_mode.mp4", "1")

main.createVideo("./test_videos/challenge_video.mp4","challenge_video_output_draw_cars.mp4","2")

main.createVideo("./test_videos/challenge_video.mp4","challenge_video_output_debug_mode_draw_cars.mp4", "3")

main.createVideo("./test_videos/project_video.mp4","project_video_output.mp4")

main.createVideo("./test_videos/project_video.mp4","project_video_output_debug_mode.mp4", "1")

main.createVideo("./test_videos/project_video.mp4","project_video_output_draw_cars.mp4", "2")

main.createVideo("./test_videos/project_video.mp4","project_video_output_debug_mode_draw_cars.mp4", "3")